Without doubt, we live in a world of viewpoints and therefore biases — therefore we cannot absolutely eliminate bias. Bias generally denotes views and opinions that are skewed through some prior experience, adherence to unreasonable perspectives that often not fact based.
Our aim is not to eliminate bias, which is futile, but to seek to gradually minimize bias.
We do so, from the perspective of machine learning and artificial intelligence. In AI and ML, we have a function called the “loss function” that algorithms seek to minimize.
Using a similar metaphor and heuristic we seek to identify features of bias and incrementally seek to identify and minimize this bias in machine learning training and predictive modeling and data.
Minimize Bias in AI 